Health outcomes in the U.S. vary dramatically across race, ethnic and social class lines, due in important part to massive disparities in human capital. Unfortunately far too little is known about the causes and potential remedies of school disengagement for disadvantaged children, perhaps because so much attention has focused on dropout itself. Yet dropping out is the endpoint of a developmental process, which begins with school disengagement earlier in the life course - that is, truancy. We propose to carry out what we believe would be the first large-scale randomized effectiveness trial designed to learn more about whether structured mentoring, monitoring, and case management can improve schooling outcomes for poor children, and to shed light on the underlying child- and family-level causes of school failure. We request funding to partner with the Chicago Public Schools (CPS) to randomly assign 120 schools to treatment and control conditions, and within schools, randomly assign around 4,000 students in 3rd to 9th grade to receive either status quo services or mentoring, monitoring and case management based on the Check and Connect (C&C) model. C&C is one of the few interventions in this literature that enjoys any support whatsoever for effects on school persistence from prior efficacy trials, although these previous trials show no significant effects on test scores or graduation (perhaps because of small sample sizes), do not examine non-school outcomes, provide limited information about mediating mechanisms, and are not informative about whether this intervention strategy can be effective at scale. Our proposed study would test this intervention and a close variant that focuses somewhat more attention on working with parents (in part to learn more about the underlying determinants of school disengagement and failure), and would carry out the study at large scale in the nation's 3rd largest city. We would use in-person surveys and administrative data to measure a wide range of different outcome domains such as violence involvement and health as well as schooling. We would also measure a range of candidate mediating processes and test different theories about the determinants of school failure by comparing intervention effects on candidate mediators and outcomes across different study sub- groups. We also propose to carry out a comprehensive benefit-cost analysis of this intervention as well. The importance of this study is self-evident from examining truancy and dropout rates in almost any major American city. The intervention is manualized, and so can be implemented at scale. Funding to launch this experiment would also enable us to track long-term outcomes at low cost using administrative data, and provides the option of carrying out follow-up studies that test the complementarities from intervening late as well as early by requesting funding in future proposals to re-randomize the sample and provide follow-up educational supports to randomly selected children from the treatment and control groups.